add our regression test suite to the repo via pytest

Closes #29

See merge request !30

add a test suite to the repo via pytest, and fix some minor bugs in _minute_gridder.py uncovered by the test rewrite

use an updated commit of glmtools that handles empty data scenes

Closes #38

See merge request !29

replace exits with raises in grid_setup, and remove the realtime check

See merge request !28

add a simple try/catch for each minute so that a single minute's failure doesn't impact other minutes

move the handling of each minute into its own function - grid_minute() - which can then be try/catch'd or potentially run in parallel

Mambaforge already has mamba installed, so this was unnecessary. In addition, it was causing builds to fail due to an incompatibility between mamba & conda versions. Related ticket: https://github.com/mamba-org/mamba/issues/1706

ensure --ctr-lat and --ctr-lon are valid latitudes & longitudes

See merge request !27

group inputs by minute and process accordingly

See merge request !26

tile processing needs to happen in the scope of each individual minute, so that we have the context to write the cspp_geo_gglm_input_files attribute correctly.

initial version of input handling where we determine minutes of interest and then work our way back to files

Allow the user to specify a --goes--sector of "meso1" or "meso2". "meso" still...

Closes #24

See merge request !25

Allow the user to specify a --goes--sector of "meso1" or "meso2". "meso" still defaults to meso1. Functionally this should only affect the output filenames.

update our dependency stack

See merge request !24

CI test prep

See merge request !21

Resolve "negative center lat/lons (meso) confuse the parser"

Closes #22

See merge request !19

small buildbucket fixes

See merge request !22

Resolve "look into GOES-18 support"

Closes #37

See merge request !20

Change the realtime flag from -1 to -r. This will allow argparse to parse negative values without extra handholding.